RESEARCH ARTICLE Open Access

Conserved roles for Polycomb RepressiveComplex 2 in the regulation of lateral organdevelopment in Aquilegia x coerulea ‘Origami’Emily J Gleason1,2 and Elena M Kramer1*

Abstract

Background: Epigenetic regulation is necessary for maintaining gene expression patterns in multicellularorganisms. The Polycomb Group (PcG) proteins form several complexes with important and deeply conservedepigenetic functions in both the plant and animal kingdoms. One such complex, the Polycomb Repressive Complex2 (PRC2), is critical to many developmental processes in plants including the regulation of major developmentaltransitions. In addition, PRC2 restricts the expression domain of various transcription factor families in Arabidopsis,including the class I KNOX genes and several of the ABCE class MADS box genes. While the functions of thesetranscription factors are known to be deeply conserved, whether or not their regulation by PRC2 is similarlyconserved remains an open question.

Results: Here we use virus-induced gene silencing (VIGS) to characterize the function of the PRC2 complex in lateralorgan development of Aquilegia x coerulea ‘Origami’, a member of the lower eudicot order Ranunculales. Leaveswith PRC2 down-regulation displayed a range of phenotypes including ruffled or curled laminae, additional lobing,and an increased frequency of higher order branching. Sepals and petals were also affected, being narrowed,distorted, or, in the case of the sepals, exhibiting partial homeotic transformation. Many of the petal limbs also hada particularly intense yellow coloration due to an accumulation of carotenoid pigments. We show that theA. x coerulea floral MADS box genes AGAMOUS1 (AqAG1), APETALA3-3 (AqAP3-3) and SEPALLATA3 (AqSEP3) areup-regulated in many tissues, while expression of the class I KNOX genes and several candidate genes involved incarotenoid production or degradation are largely unaffected.

Conclusions: PRC2 targeting of several floral MADS box genes may be conserved in dicots, but other knowntargets do not appear to be. In the case of the type I KNOX genes, this may reflect a regulatory shift associatedwith the evolution of compound leaves.

Keywords: Polycomb repressive complex 2 (PRC2), Compound leaves, AGAMOUS, Class I KNOX genes,Carotenoid biosynthesis, Epigenetics, Evolution, Aquilegia

BackgroundMaintenance of proper gene expression in differentiatedcells is essential for the development of multicellularorganisms and epigenetic regulation is an importantplayer in this process Reviewed in: [1-3]. One family ofproteins with deeply conserved functions in epigeneticregulation is the Polycomb Group (PcG). The PcG was

first discovered in Drosophila melanogaster as repressorsof the HOX genes [4]. Several PcG complexes exist inboth plants and animals, each with distinct functions inepigenetic silencing Reviewed in: [5,6]. However, onlythe Polycomb Repressive Complex 2 (PRC2) has beenwell characterized in multiple plant models Reviewed in:[5,7]. The main function of the PRC2 complex istrimethylation of lysine 27 of histone H3 (H3K27), ahistone modification known to suppress gene expression[8]. The PRC2 contains four core proteins; the histonemethyltransferase Enhancer of Zeste (E(z)), and threeother proteins thought to enhance PRC2 binding to

* Correspondence: ekramer@oeb.harvard.edu1Department of Organismic and Evolutionary Biology, Harvard University, 16Divinity Ave., Cambridge, MA 02138, USAFull list of author information is available at the end of the article

© 2013 Gleason and Kramer; licensee BioMed Central Ltd. This is an open access article distributed under the terms of theCreative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use,distribution, and reproduction in any medium, provided the original work is properly cited.

Gleason and Kramer BMC Plant Biology 2013, 13:185http://www.biomedcentral.com/1471-2229/13/185

nucleosome [9]. These include Suppressor of Zeste 12(Su(z)12) and Extra Sex Combs (ESC), known respectivelyas EMBRYONIC FLOWER 2 (EMF2) and FERTILIZATIONINDEPENDENT ENDOSPERM (FIE) in plants, andMulti-Copy Suppressor of IRA 1 (MSI1) Reviewed in: [10].The E(z) lineage in plants has experienced an ancientduplication such that most angiosperms have at least twoparalogs, known as CURLY LEAF (CLF) and SWINGER(SWN) [11]. Many plant species have additional duplica-tions in the core PRC2 loci that allow them to formseveral PRC2 complexes often with distinct developmentalfunctions [12,13].PRC2 is involved in a number of important develop-

mental transitions. In the plant model system A. thaliana,these functions include endosperm development, earlyrepression of flowering to allow proper vegetative devel-opment, the eventual transition to flowering, and flowerorganogenesis [14-17]. In grasses, the PRC2 complex playsroles in floral induction (rice and barley), flower develop-ment (rice), suppressing cell divisions in the unfertilizedovule (rice), and endosperm development (rice and maize)[12,18,19]. In the moss model Physcomitrella patens,PRC2-dependent remodeling appear to be required forthe switch from gametophyte to sporophyte development[20,21].In addition to its role in developmental transitions,

PRC2 has been suggested to function in lateral organdevelopment in A. thaliana. In fact, the first descriptionof a plant PRC2 function was discovered with thecharacterization of the clf mutant in A. thaliana [17]. Theclf plants had severely curled leaves, smaller narrowersepals and petals, and partial homeotic transformations ofsepals and petals towards carpel and stamen identity,respectively. Two MADS box genes, the C class memberAGAMOUS (AG) and the B class representativeAPETALA3 (AP3) were shown to be over-expressed in clfmutants, suggesting that the PRC2 complex was requiredfor stable repression of these genes [17]. This was particu-larly interesting because MADS box genes regulatehomeotic floral organ identity in plants somewhat analo-gously to the way HOX genes regulate segment identity inanimals [22-25]. Further studies have subsequently shownthat the E class MADS SEPALLATA3 (SEP3) is similarlyup-regulated in clfmutants [26]. PRC2 has also been shownto regulate the expression of the class I KNOX genesduring vegetative development. The class I KNOX genesare a family of homeobox domain-containing loci in plantsthat have conserved roles in promoting pluripotency in theshoot apical meristem and in compound leaf development[27,28]. Katz et al [29] found that in addition to thephenotypes reported in clf mutant plants, FIE cosuppressedplants also had loss of apical dominance and fasciatedstems, rolled leaves with varying degrees of serration, lossof phyllotaxy in the inflorescence, and many problems with

ovary and ovule development. They further demonstratedthat several class I KNOX genes, including BREVIPEDI-CELLUS (BP), KNOTTED-LIKE FROM ARABIDOPSISTHALIANA 2 (KNAT2), and SHOOTMERISTEMLESS(STM), were over-expressed in rosette leaves of FIE silencedplants. In clf mutants, STM and KNAT2 were over-expressed but BP was not, possibly because the CLFparalog SWN was acting redundantly. The class I KNOXgenes MOSS KNOTTED1-LIKE 2 and 5 (MKN2 andMKN5) were also shown to be over-expressed in PpFIEmutant gametophytes [21,30], suggesting that PRC2targeting of the class I KNOX genes may be deeplyconserved.While the functions of the floral ABC class and type I

KNOX genes are thought to be conserved across angio-sperms, comparative studies of their regulation have largelyfocused on upstream transcription factors, such as LEAFYor ARP family members [31,32]. In order to begin address-ing the question of whether PRC2-targeting interactionsare similarly conserved, we have examined the functions ofPRC2 members in lateral organ development of theemerging model system Aquilegia. The genus Aquilegia is amember of an early diverging lineage of the eudicotyledo-nous flowering plants, the Ranunculales, that arose beforethe radiation of the core eudicots Reviewed in: [33]. Ittherefore can be used as a rough phylogenetic midpoint be-tween A. thaliana and model systems in the grasses [34].Additionally, many ecological, evolutionary and geneticstudies have been conducted in Aquilegia over the past50 years. These have taken advantage of its small genome(n = 7, approximately 300 Mbp) as well as a number ofmore recent genomic tools, including the fully sequencedAquilegia x coerulea genome (http://www.phytozome.net/search.php?method = Org_Acoerulea) Reviewed in: [33,35].The reverse genetic tool virus-induced gene silencing(VIGS) has been optimized in several species of Aquilegia[36] for both leaf and floral development [37-40]. Previouslywe examined the evolution and expression of the PRC2family in Aquilegia [41] and found that the genomecontains a simple complement of PRC2 homologs: onecopy each of the two plant E(z) homologs, AqCLF andAqSWN; an ESC homolog, AqFIE; a Su(z)12 homolog,AqEMF2; and a copy of MSI1, AqMSI1. We initiallyassessed gene expression throughout Aquilegia vulgarisdevelopment due to its strong vernalization dependencyand found no obvious tissue or stage specialization.Furthermore, the ancient paralogs, AqCLF and AqSWN, arenot imprinted in Aquilegia endosperm as is seen in otherplant species [19,41].In the current study we have used VIGS to knock

down the expression of AqFIE [Genbank: JN944599] andAqEMF2 [Genbank: JN944598] in unvernalized andvernalized Aquilegia coerulea ‘Origami’ plants using theANTHOCYANIN SYNTHASE (AqANS) as a marker

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 2 of 13http://www.biomedcentral.com/1471-2229/13/185

gene. Due to limitations of the VIGS approach, it is notpossible to assess many life cycle transitions, most not-ably flowering time, but lateral organ development canstill serve as a useful model for PRC2 function. We findthat PRC2 plays a role in leaf and floral organ develop-ment in A. x coerulea, particularly via down-regulationof the floral MADS box genes. This has allowed us toidentify PRC2 targets that appear to be conservedbetween Arabidopsis and Aquilegia as well as somenovel PRC2-regulated pathways.

MethodsVirus-induced gene silencingThe Aquilegia VIGS protocol was preformed as describedpreviously [36]. TRV2-AqCLF-AqANS, TRV2-AqSWN-AqANS, TRV2-AqFIE-AqANS and TRV2-AqEMF2-AqANSconstructs were prepared by PCR amplifying approxi-mately 300 bp regions of each gene using primers thatadded EcoR1 and XbaI restriction sites to the 5′ and 3′ends of the PRC products (see Additional file 1). The PCRproducts were then purified and cloned into the TRV2-AqANS construct [36] and electroporated into Agrobacter-ium strain GV101. A. x coerulea seedlings were grown toapproximately the 4 to 6 leaf stage and then either treatedas described in Gould and Kramer [36] for unvernalizedsamples or as described in Sharma and Kramer [37] forplants that had been vernalized for approximately 4 weeksat 4°C [36,37]. The TRV2-AqANS and TRV2-AqFIE-AqANS constructs were each used to treat approximately400 plants over 4 rounds of VIGS. Approximately 250 ofthese plants were VIGS treated before vernalization andapproximately 150 of these plants were treated aftervernalization. The TRV2-AqEMF2-AqANS construct wasused to treat approximately 100 plants; roughly 50 of theseplants were treated before vernalization and 50 weretreated after vernalization. Leaves, petals, and sepals show-ing AqANS silencing were photographed, collected, andstored at -80°C for RNA analysis.

RT-PCRRNA was extracted from control (AqANS silenced) andexperimental (AqFIE and AqEMF2 VIGS-treated) tissue.One half of the AqANS silenced (control) leaves werefrom separate unvernalized plants (C1 and C2) and halfwere from separate vernalized plants (C3 and C4). Five ofthe TRV2-AqFIE-AqANS treated leaves were from separ-ate unvernalized plants (F2, F4, F5, F7, and F8) whilethree were collected from separate vernalized plants(F1, F3, and F6). All of the TRV2-AqEMF2-AqANS leaveswere collected from separate vernalized plants (E1-E4).Sample numbers do not indicate order of leaf appearancebut were collected at roughly the same stages of develop-ment. We selected a variety of observed phenotypes foreach set of samples. For leaves, the RNeasy Mini Kit

(Qiagen, Valencia, CA) was used. For petals and sepalsRNA was extracted using the Pure-Link Plant RNAReagent small scale RNA isolation protocol (Ambion,Austin, TX). RNA was treated with Turbo DNase (Ambion,Austin, TX) and cDNA was synthesized from 1 μg of totalRNA using Superscript II reverse transcriptase (Invitrogen,Carlsbad, CA) and oligo (dT) primers. cDNA was diluted1:5 prior to use.Amplification was performed using AccuStart PCR

SuperMix (Quanta Biosciences Inc, Gaithersburg, MD).The amplification program began with 1 minute activationstep at 94°C, followed by a 20 second denaturing step at94°C, a 15 second annealing step at 55°C, and a 15 secondextension at 72°C, repeated for 30 cycles. This cyclenumber was chosen for optimal detection of AqFIE andAqEMF2, which are expressed at relatively low levels inmature organs, especially compared to the high expressionlevels of AqIPP2. All primers used are listed in Additionalfile 1. Amplification of ISOPENTYL PYROPHOSPHATE:DIMETHYLALLYL PYROPHOSPHATE ISOMERASE2(AqIPP2) was used as a positive control [38,42]. To test forexpression of APETALA3-1 (AqAP3-1), APETALA3-2(AqAP3-2), APETALA3-3 (AqAP3-3), and FUL-like- 1(AqFL1) in VIGS-treated leaves, cDNA from severalleaves were pooled together prior to amplification. Thecontrol pool consisted of AqANS-silenced control leavesC1-4, the AqFIE VIGS-treated pool consisted of AqFIEleaves F3-6, and the AqEMF2 VIGS-treated pool con-sisted of AqEMF2 leaves E1-4.

qRT-PCRcDNA was prepared from VIGS-treated tissue asdescribed above. For the carpel sample, carpels werecollected from 3 anthesis stage wild type plants andpooled together. RNA was extracted using the RNeasyMini Kit (Qiagen, Valencia, CA) and treated asdescribed above. cDNA from VIGS-treated tissue wasthen pooled together and diluted 1:10. The controlsepal pool consisted of AqANS-silenced control sepalsC1-4, the control petal pool consisted of AqANS-silenced control petals C1-4, the AqFIE sepal poolconsisted of AqFIE VIGS-treated sepals F2, 3, 5, and6, the AqFIE petal pool consisted of AqFIE VIGS-treated petals F2, 3, 5, and 6, the AqEMF2 sepal poolconsisted of AqEMF2 VIGS-treated sepals E2, 3, and4 s, and the AqEMF2 petal pool consisted of AqEMF2VIGS-treated petals E1 and E2. qRT-PCR was per-formed using PerfeCTa qPCR FastMix, Low ROX(Quant Biosciences Inc., Gaithersburg, MD) in theStratagene Mx3005P QPCR system to study the relativeexpression of AqAG1 and AqAG2. AqIPP2 expressionwas used for value normalization. All primers are listedin Additional file 1.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 3 of 13http://www.biomedcentral.com/1471-2229/13/185

MicroscopyPetals from wild type, AqANS VIGS-treated, and AqEMF2VIGS-treated plants were stored at −80°C and thenwarmed to room temperature and mounted whole on glassslides in water. Cells were visualized in the Harvard Centerfor Biological Imaging on a Zeiss AxioImager Z2 micro-scope using trans-illumination with white light. Imageswere taken using a Zeiss AxioCam Mrc digital camera.

ResultsWe treated both unvernalized and vernalized plants withTRV2 constructs containing either AqANS-AqFIE orAqANS-AqEMF2 fragments. TRV2-AqANS treated plantswere used as controls throughout. Phenotypes of AqFIEand AqEMF2 silenced plants were equivalent and will bediscussed together. We also treated a small number ofunvernalized plants with AqANS-AqCLF and AqANS-AqSWN VIGS constructs. Phenotypes from these plantswere similar to those seen in AqFIE and AqEMF2, butwere weaker (data not shown), most likely due to partialredundancy between AqCLF and AqSWN. Thus we choseto focus on AqFIE and AqEMF2 VIGS-treated tissue. Asis common for VIGS-treated plants, we recovered a rangeof phenotypes in a small percentage of VIGS-treatedplants (roughly 10-15% of plants in each round) [36]. Inthe current experiment there is the added component thatphenotypes are likely due to mis-expression of PRC2target genes, and are therefore likely to have an addedcomplexity due to ectopic expression of a potentially widerange of target loci.

Vegetative phenotypesWild type Aquilegia leaves are compound, typically bear-ing three leaflets that are themselves divided into two tothree lobes (Figure 1A). Although these leaflets are oftenrelatively deeply lobed, they do not generally produceelongated, higher order petiolules within the leaflets.However, A. x coerulea does display heteroblasty over thecourse of its lifespan, varying leaf morphology as theindividual progresses from the vegetative to the repro-ductive stage (Additional file 2). In late reproductiveadult stages, higher order petiolules may be observed inwhich the central lobe of each leaflet becomes itselfa separate leaflet borne its own petiolule (Additionalfile 2C). Using the terminology of Kim et al. 2003 [32], allof these leaf forms are non-peltately palmate in that theleaflets are not radially positioned around the terminusof the primary petiole.We observed AqANS silencing in 10-15% of treated

plants across the AqFIE- and AqEMF2-VIGS experi-ments. In addition to the AqANS-silencing, the leaves ofthese plants showed a complex set of phenotypes. Themost consistently observed perturbation was curled orruffled laminae that typically curled toward the abaxial

surface (~10-12% of treated plants and, thus, the majorityof silenced plants) (Figure 1F, H, J-L). We also observedan increased frequency of higher order branching inwhich fully formed petiolules developed within theleaflet, creating as many as ten or twelve distinct leafletsrather than the usual three (Figure 1B-F, H, L andAdditional file 2E). While we have never observed suchhigher order branching in control leaves, either in thecontext of these experiments or others [40], we obtained15 leaves from a total of 10 AqFIE- and AqEMF2-treatedplants that exhibited increased branching. When quanti-fied (Additional file 2E), the presence of higher orderpetiolules is significant at p < 0.05 for unvernalized lateralleaflets but not significant for the other stages/leaflettypes. However, it is obvious that there is much morebranching variation in silenced leaflets than in controls.In many cases, the margins of the laminae had additionallobing relative to control leaves (observed in 10% oftreated plants) (Figure 1B-E, K) and, in a small numberof cases, the central lobe of the terminal leaflet wasseverely reduced (seen in multiple leaves from 5 plants)(Figure 1D, M). Laminar area was highly variable withsome leaflets appearing to have expanded area (~25plants) (Figure 1F) while others seemed reduced (~8plants) (Figure 1G, I, M). In two plants, ectopic finger-like projections were observed on the adaxial surface oflaminae (Figure 1F), which was never observed in controlleaflets.

Floral phenotypesWild type A. x coerulea flowers possess five organ types:sepals, petals, stamens, staminodia and carpels [39]. Wehave focused on the sepals and petals because theyshowed strong phenotypes in the silenced flowers. Wildtype sepals are flat and ovate with an entire margin(Figure 2A-B). The petals are notable for the presence ofa long hollow nectar spur, which forms near the attach-ment point (Figure 2A). This feature divides the organinto two regions, the proximal spur and the distal limb.Spurs in A. x coerulea are typically 5-6 cm in length andslightly curved. The limb region is relatively flat with arounded, weakly lobed margin (Figure 2C). In 25 flowersfrom vernalized AqFIE- and AqEMF2-treated plants, weobserved sepals that were narrower than wildtype organsand dramatically folded towards the adaxial surface(Figure 2D, F, L, P). In severely affected flowers, petalswere narrowed and stunted (10 flowers, Figure 2D, G, Q)or exhibited sharply bent spurs (12 flowers, Figure 2H-I,K, M, Q). In two AqEMF2-silenced flowers, the sepalsexhibited chimeric petal identity including ectopic spurformation (Figure 2M-N). Perhaps most surprising, manyof the perianth organs had a definite yellow hue, with thepetal limbs showing particularly intense yellow coloration(observed in at least one organ from 15 flowers) (Figure 2E,

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 4 of 13http://www.biomedcentral.com/1471-2229/13/185

I-M, O). Such coloration was not observed in AqANS-silenced control flowers (Figure 2A-C). Examination of theAqFIE- and AqEMF2-silenced organs under high magnifi-cation reveals that yellow pigment is deposited in plastids(Additional file 3A), consistent with carotenoids ratherthan the vacuole-based aurones that are produced in someAquilegia species [43,44].

Assessment of AqFIE and AqEMF2 down-regulationDue to limited RNA availability, we used standardRT-PCR to assess target gene down-regulation in leaves,sepals and petals compared to their expression inAqANS silenced control tissue. Even in AqANS silenced

(control) tissue, AqFIE and AqEMF2 are expressed atlow levels relative to the loading control AqIPP2. Theexperimental samples (F1-8 and E1-4) were selected torepresent all of the observed phenotypes and werederived from separate plants. This analysis demonstratedthat in the TRV2-AqFIE-AqANS treated plants, AqFIEwas strongly down-regulated, being undetectable in anumber of samples (Figure 3A and Figure 4). Likewise,AqEMF2 expression is reduced to undetectable levels inmost tested AqEMF2-silenced samples (Figure 3A andFigure 4). We also tested for AqEMF2 in AqFIE-treatedplants and vice versa, and found that AqEMF2 levels areoften reduced in AqFIE-treated leaves, although the

Figure 1 Vegetative phenotypes of PRC2 VIGS-treated plants. A. AqANS-treated leaf (Control) with three lobed leaflets. First order petiolulesare marked with asterisks. B-I. AqFIE-silenced leaves and leaflets (abbreviated aqfie). B. Entire leaf with highly branched leaflets. C-E. Each leafletfrom the leaf shown in B with higher order petiolules marked with asterisks and reduced central lobe indicated with an arrow. Leaflets arearranged in clockwise order starting with the left lateral leaflet in B. F. Leaflet with curled laminae, increased branching (asterisks) and ectopicoutgrowth on the adaxial lamina (white arrowhead). G. Leaflet with reduced lamina and narrow lobes that are deeply divided. H. Entire leafshowing increasing internal branching (asterisks) and curling. I. Entire leaf with deep lobes and aberrantly shaped laminae. J-M. AqEMF2-silencedleaves (abbreviated aqemf2). J. Entire leaf showing curled/ruffled laminae and deep lobing. K. Central leaflet from J exhibiting curled laminae,increased degree of lobing and serration. L. Entire leaf with internal branching (asterisks) and curled laminae. M. Leaflet with reduced central lobe(arrow). Scale bars: 1 cm.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 5 of 13http://www.biomedcentral.com/1471-2229/13/185

Figure 2 Floral phenotypes of PRC2 VIGS-treated plants. A-C. AqANS-silenced control flower and perianth organs (Control). A. Entire flower.B. Entire sepal. C. Petal limb. D-J. AqFIE-silenced flowers and organs (abbreviated aqfie). D. Severely affected flower. E. Moderately affected flower.F. Narrow, folded sepal of flower in D. G. Narrow, stunted petal of flower in D. H-I. Petals with bent spurs from moderately affected flowers.J. Yellow limb of moderately affected petal. K-Q. AqEMF2-silenced flowers and organs (abbreviated aqemf2). K-L. Severely affected flowers.M-N. Sepal/petal chimeras from first whorl of flowers such as K. O. Yellow limb of second whorl petal from flower in K. P. Narrow, folded sepalfrom flower in L. Q. Narrow, bent petal from flower in L. Scale bars: 1 cm.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 6 of 13http://www.biomedcentral.com/1471-2229/13/185

reciprocal is generally not true (Figure 3A). Furthermore,we tested the other PRC2-complex members, AqCLF andAqSWN, and found no consistent evidence of their down-regulation in either type of silenced tissue (Additionalfile 4A).

Assessment of candidate gene expressionWe tested for ectopic expression of a wide panel ofpotential target genes, with a focus on the floral organidentity loci and type I KNOX homologs (Figure 3,Figure 5, and Additional file 4B). We again compared

Figure 3 RT-PCR expression data in PRC2 VIGS-treated leaves. AqIPP2 was used as a loading control for all reactions. Note that theexpression of AqFIE and AqEMF2 are low relative to the expression of AqIPP2. A. Expression of AqFIE and AqEMF2 in AqANS-silenced control leaves(C1-C4), and AqFIE-silenced (F1-F8) and AqEMF3-silenced (E1-E4) leaves. AqFIE is clearly down-regulated in AqFIE-silenced tissue and, likewise,AqEMF2 is down-regulated in AqEMF2-silenced tissue. Interestingly, AqEMF2 also appears to be down-regulated in AqFIE-treated leaves but AqFIEexpression is unaffected in AqEMF2-treated leaves. B. Expression of several floral organ identity genes in AqANS-silenced control leaves (C1-C4),and AqFIE-silenced (F1-F8) and AqEMF3-silenced (E1-E4) leaves. In several of the AqFIE down-regulated leaves and all of the AqEMF2down-regulated leaves, AqAG1 is over-expressed compared to AqANS-silenced control leaves. While the expression of the SEPALLATA homologs isvariable in both control and experimental leaves, AqSEP3 may be up-regulated in some of the AqFIE- and all of the AqEMF2-silenced leaves.C. Expression of several of the A. x coerulea class I KNOX genes in AqANS-silenced control leaves (C1-C4), and AqFIE-silenced (F1-F8) andAqEMF3-silenced (E1-E4) leaves. Expression of these genes is unaffected in the mature AqFIE- and AqEMF2-silenced leaves.

Figure 4 Expression of AqFIE and AqEMF2 in VIGS-treated floral organs. A. Expression of AqFIE and AqEMF2 in AqANS-silenced control sepals(C1-C4), and AqFIE- (F2-F6) and AqEMF3- (E1-E4s/p) treated first whorl organs. AqFIE is down-regulated in all of the AqFIE-treated sepals. Likewise,AqEMF2 is down-regulated in AqEMF2-treated first whorl organs. Unlike the pattern in leaves, AqEMF2 is not down-regulated in AqFIE-treatedsepals. B. Expression of AqFIE and AqEMF2 in AqANS-silenced control petals (C1-C4), and AqFIE- (F1-F6) and AqEMF2- (E1 and E2) treated petals.AqFIE is down-regulated in all of the AqFIE-treated petals while AqEMF2 is down-regulated in E1 and also in F3.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 7 of 13http://www.biomedcentral.com/1471-2229/13/185

the expression of these genes to expression in AqANS-silenced control tissue. One of the two A. x coerulea theC class MADS box genes, AGAMOUS 1 (AqAG1) (seeAdditional file 1 for all gene identification numbers), isconsistently up-regulated in silenced leaves and floralorgans. The second AGAMOUS homologs, AGAMOUS2(AqAG2) may also be slightly up-regulated in some ofthe leaves, although AqAG2 shows basal expression incontrol floral organs (Figures 3 and 5). The three A. xcoerulea SEPALLATA paralogs (AqSEP1, AqSEP2, andAqSEP3) are somewhat difficult to assess because they

are variably expressed in control leaves but AqSEP3 inparticular seems to be up-regulated in AqEMF2-silencedleaves (Figure 3B). These genes were not assessed in floralorgans because they are already broadly expressed in thesetissues. A. x coerulea also has three paralogs of the B classMADS box gene, APETALA3 (AqAP3-1, AqAP3-2, andAqAP3-3). The petal-specific AqAP3-3 locus is highly up-regulated in AqEMF2-silenced sepals, which also showedchimeric sepal/petal identity in several cases (Figure 5A).Additionally two of the three AP3 paralogs are moderatelyup-regulated in PRC2 VIGS-treated leaves (Additional

Figure 5 Expression of candidate genes in PRC2 VIGS-treated floral organs. A. Expression of several floral organ identity genes inAqANS-silenced control sepals (C1-C4), and AqFIE- (F2-F6) and AqEMF2- (E1-E4s/p) treated first whorl organs. AqAG1 is up-regulated in allAqFIE- and AqEMF2-treated organs compared to the controls.AqAP3-3 also appears to be up-regulated in some of the sepals, particularly inAqEMF2 down-regulated first whorl organs, several of which were in fact sepal/petal chimeras (s/p). Expression of AqAP3-2 and AqFL1 is variable inmature sepals and is difficult to assess. AqAG2 and AqAP3-1 expression does not appear to be affected in silenced tissue. B. Expression of severalfloral organ identity genes in AqANS-silenced control petals (C1-C4), and AqFIE- (F2-F6) and AqEMF2- (E1-E4s/p) VIGS-treated petals. AqAG1 isup-regulated in all AqFIE- and AqEMF2-treated tissue compared to the controls. C and D. Quantitative Real Time PCR analysis of expression ofAqAG1 and AqAG2 in AqFIE, AqEMF2, and AqANS control silenced tissue and wild type carpels. cDNA from two to four samples was pooledtogether prior to analysis. For each data point, three technical replicates were analyzed. AqIPP2 expression was used for normalization. C. Averagefold change in the expression of AqAG1 in AqFIE, AqEMF2, and AqANS control silenced tissue normalized to wild type carpels with SD error bars.D. Average fold change in the expression of AqAG2 in AqFIE, AqEMF2, and AqANS control silenced tissue normalized to wild type carpels with SDerror bars.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 8 of 13http://www.biomedcentral.com/1471-2229/13/185

file 4B), but the expression of AqAP3-1 and AqAP3-2is unaffected in mature sepals and petals (Figure 5Aand B). We also looked at the expression of FUL-like 1(AqFL1), which is normally expressed in early leaves,but no ectopic expression was detected (Figure 5 andAdditional file 4B).Next, we tested for up-regulation of three of the five

A. x coerulea class I KNOX genes. No significant ectopicKNOX gene expression could be detected in theleaves. Weak expression of the SHOOTMERISTEMLESS2 (AqSTM2) and KNOTTED (AqKN) homologs isdetected in AqEMF2-silenced leaves (Figure 3C), however,we also occasionally detected comparable expression ofthese genes in control (AqANS-silenced) leaves, so it isdifficult to ascribe significance to this expression. Giventhe lack of clear up-regulation in leaves and due to alimited amount of floral RNA, class I KNOX gene expres-sion was not tested in the floral organs.Although AqAG1 is consistently over-expressed in

AqFIE and AqEMF2 silenced sepals and petals, we neversaw any evidence of carpel identity in these organs. Wetherefore pooled cDNA from several AqANS- (control),AqFIE-, and AqEMF2-treated petals and sepals and usedqRT-PCR to further examined the expression of AqAG1and AqAG2 in these organs as well as in wild typecarpels (Figure 5C and D). We found that while AqAG1was clearly up-regulated in AqFIE and AqEMF2 silencedorgans compared to the controls, AqAG1 expression wasstill much lower than in wild type carpels (about 0.05 to0.2 fold). In contrast, AqAG2 expression was similar incontrol and PRC2 silenced tissue, but much lower thanin wild type carpels.Lastly, in an effort to investigate the carotenoid

production, we identified the likely A. x coeruleahomologs of a range of components of the carotenoidpathway in A. thaliana, including enzymes involvedin production (PHYTOENE SYNTHASE (PSY) andCAROTENOID ISOMERASE (CRTISO)) and break-down (CAROTENOID CLEAVAGE DIOXYGENASE 4(CCD4) and 9-CIS-EPOXYCAROTENOID DIOXYGEN-ASE 3 (NCED3)) of carotenoids [45]. A. x coerulea hastwo copies CCD4 (AqCCD4 and AqCCD4L) and twogenes that are closely related to A. thaliana PSY (AqP-SYL1 and AqPSYL2). Previous studies in A. thalianahave indicated that both CRTISO and NCED3 are posi-tively epigenetically regulated by other SET domaincontaining proteins so we were particularly interestedin the expression of these genes in AqFIE and AqEMF2down-regulated tissue [46,47]. We used RT-PCR toexamine the expression of these six genes in AqANS-(control), AqFIE-, and AqEMF2-treated petals (Add-itional file 3B). Given the observed phenotypes, wemight expect the expression of AqPSYL1, AqPSYL2, orAqCRTISO to be up-regulated or AqCCD4, AqCCD4L, or

AqNCED3 to be down-regulated. Unfortunately, no clearpatterns are apparent from these reactions.

DiscussionAqFIE and AqEMF2 VIGS-treated plants displayed arange of lateral organ phenotypes. Silenced leaves oftenhad ruffled or curled lamina, additional lobing, and anincreased frequency of higher order branching. The peri-anth organs were generally narrower than wild typeorgans. Sepals were also curled and petals were stunted orhad bent spurs, while petal limbs also had a particularlyintense yellow coloration seemingly due to an accumula-tion of carotenoid pigments in these cells. Many of thephenotypes we observed are similar to those seen in clfmutants and FIE cosuppressed A. thaliana, includingcurled leaves and narrow perianth organs [17,29]. Unlikeclf mutants and AG over-expressers in A. thaliana,dramatic transformation towards carpel identity was notobserved in the AqFIE- and AqEMF2-treated sepals orpetals. However, the level of AqAG1 expression in theseorgans was much less than what is seen in wild type Aqui-legia carpels. Interestingly, the distinct folded morphologyof the sepals may suggest slight transformation towardscarpel identity as silenced leaves were folded towards theabaxial surface while the sepals were dramatically foldedtowards the adaxial surface, which is similar to the foldingpattern of the Aquilegia carpel [48].It is interesting to note that in AqFIE silenced leaves,

AqEMF2 is also down-regulated. The reverse is not truein AqEMF2 silenced leaves, and AqEMF2 expression isnot affected in AqFIE silenced floral organs. This resultsuggests that PRC2 may be directly or indirectly regulat-ing AqEMF2 expression in A. x coerulea leaves, whichcould account for the generally more severe phenotypesobserved in AqFIE silenced leaves compared to AqEMF2silenced leaves. AqEMF2 is the only member of the com-plex that appears to be PRC2-regulated as the expressionof AqCLF and AqSWN is not affected in PRC2 down-regulated leaves. In general, the potential for this type ofcross-regulation is relatively unexplored in A. thalianaand, therefore, bears further study.In our analysis of candidate target genes, we found that

AqAG1 is often ectopically expressed in PRC2 down-regulated tissue. AqAP3-3 and AqSEP3 are also up-regulated in some organs, but expression of the class IKNOX genes and several candidate genes involved incarotenoid production or degradation seem largelyunaffected. Mutations in AG and SEP3 are known tosuppress the curled leaf phenotype in clf mutant plantswhile over-expression of these MADS box genes, whichthemselves function together in a complex [49], is thoughtto be the cause of the curled leaf phenotype [26]. It is,therefore, possible that over-expression of AqAG1 andAqSEP3 is similarly responsible for many of the observed

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 9 of 13http://www.biomedcentral.com/1471-2229/13/185

phenotypes in AqFIE and AqEMF2 silenced leaves. Thesefindings lead us to conclude that PRC2-based regulationof AG and SEP3 homologs is deeply conserved in eudicots.It has recently been shown that several chromatin remod-eling factors associate with MADS complexes and onemodel is that an important function of MADS domaincomplexes may be to recruit chromatin remodelingcomplexes to target loci in order to alter transcription ofthese genes and direct organ development [50,51]. Forexample, RELATIVE OF EARLY FLOWERING 6 (REF6)was enriched in protein complexes that were isolated viaimmunoprecipitation using tagged ABCE class MADSbox proteins [50]. REF6 has been shown to specificallydemethylate H3K27me3, the histone modification depos-ited by PRC2 [52]. Activation of SEP3 by APETALA1(AP1) in A. thaliana results in the reduction of H3K27me3at the SEP3 promoter, suggesting that AP1 may recruitREF6 to the SEP3 promoter in order to help induce SEP3gene function [50]. Our data suggests that this key depend-ency on epigenetic regulation for the switch from vegeta-tive to floral development may be important outside of A.thaliana. There are some complications, however. Of thetwo A. x coerulea AG homologs, only one, AqAG1, isstrongly regulated by PRC2. Perhaps consistent with thisobservation, sequencing of the Aquilegia genome (http://www.phytozome.net/search.php?method=Org_Acoerulea)reveals that AqAG1 does contain the large regulatorysecond intron that is common to AG homologs [53,54]while AqAG2’s second intron is much smaller. Theseresults suggest that PRC2 regulation can be directed in aparalog-specific fashion and may even play some role inthe distinct expression patterns observed among thesegene copies [39].The class I KNOX genes are directly or indirectly

regulated by PRC2 in both A. thaliana and Physcomi-trella, however, we detected little or no increase inKNOX gene expression in our AqFIE and AqEMF2silenced leaves. This is somewhat surprising because ofthe higher order branching that we observed in si-lenced leaves, including several of the tested RNA sam-ples. The class I KNOX genes are thought to play arole in compound leaf development in a number ofspecies. In many, but not all, compound leafed taxawhere KNOX gene expression has been studied, in-cluding Aquilegia, it has been shown that the genes areexpressed in the shoot apical meristem and down-regulated in incipient leaf primordia (P0), butsubsequently turned back on in early leaf primordia[28]. Down-regulation of class I KNOX genes in theleaves of models such as tomato or Cardamine causesreduced branching while over-expression leads to in-creased branching [55,56], suggesting that KNOX genesact to maintain indeterminacy in compound leaves andpromote leaflet initiation.

There are several possible explanations for why wedid not observe significant ectopic KNOX gene expres-sion in our VIGS-treated leaves. First, it is possible theKNOX genes were ectopically expressed early in leafdevelopment when the higher order branching actuallydeveloped, but were later down-regulated by redundantmechanisms, such as ASYMMETRIC LEAVES 1 (AS1)-mediated repression [57,58]. In A. thaliana AS1 medi-ated silencing of some of the KNOX genes has beenshown to require the PRC2 complex and it is thoughtthat AS1 and AS2 directly recruit the PRC2 complex toKNOX loci [59]. However, it is important to rememberthat in other taxa with compound leaves, the KNOXand AS1 homologs have lost their mutually exclusiveregulatory interactions and are expressed togetherat later stages [32]. This may suggest that the AS1-dependent epigenetic silencing of KNOX genes thathas been described in several simple-leafed models[57,58] does not hold for plants with compound leaves.Along these lines, it is also possible that the increasedbranching phenotypes are due to other factors, such asaccelerated phase change or novel genetic mechanismsregulating leaflet branching in Aquilegia. For instance,a recent functional study of the gene AqFL1 in A.x coerulea revealed that it promotes proper leaf margindevelopment, a unique finding for homologs of thisgene lineage [40]. This raises the possibility that factorsother than the KNOX genes contribute to compoundleaf branching in Aquilegia.In addition to the conserved role in regulating AG,

AP3, and SEP3, A. x coerulea PRC2 may target novelpathways, including those regulating carotenoid pro-duction or degradation. In A. thaliana patches of yel-low anther-like tissue are observed on clf mutant petals[17]. However, the yellow pigmentation we observed isdue to the accumulation of carotenoids in the plastidsrather than to a partial homeotic transformation. Whilegenes in the carotenoid pathway are not known to besuppressed by PRC2, some loci are positively epigeneti-cally regulated in A. thaliana. Previous studies haveshown that a major enzyme in the carotenoid biosyn-thesis pathway, CRTISO, requires the chromatin modi-fying enzyme SET DOMAIN GROUP 8 (SDG8) tomaintain its expression [46]. NCED3, an enzyme thatcleaves some types of carotenoids as a part of abscisicacid (ABA) synthesis, is similarly epigenetically regu-lated by the A. thaliana trithorax homolog ATX1 [47].While none of the genes we tested were consistentlyup- or down-regulated in AqFIE and AqEMF2 silencedpetals, carotenoid production is very genetically com-plex and we were unable to test all of the candidateloci [60]. Thus, it seems likely that PRC2 regulatesan as yet unidentified enzyme in this pathway in A.x coerulea.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 10 of 13http://www.biomedcentral.com/1471-2229/13/185

Conclusions

� A critical role for PRC2 in maintaining therepression of AG, SEP3, and possibly AP3 appears tobe conserved across eudicots. This conservationunderscores the importance of chromatinremodeling factors in regulating the floral transitionand the proper localization of floral organ identity.

� Class I KNOX genes are not ectopically expressed inPRC2 down-regulated tissue in A. x coerulea,possibly due to a regulatory shift associated with theevolution of compound leaves.

� A. x coerulea PRC2 plays a significant role inregulating the carotenoid pathway in floral organs,which has not been observed in other taxa.

� This study, the first to examine PRC2 function inangiosperms outside A. thaliana or the grasses,highlights how little we still know about the generalconservation or targeting mechanisms underlyingPRC2 function in major developmental transitions.

Additional files

Additional file 1: Table of all PCR primers.

Additional file 2: Heteroblasty in A. x coerulea leaves. A.Unvernalized leaf with 3 major lobes in the lateral leaflets. B. Unvernalizedleaf with 2 major lobes in the lateral leaflets. C. Vernalized leaf withhigher order petiolules where the central lobe of each leaflet is aseparate leaflet borne on a petiolule (asterisks). D. Vernalized leaf with 2major lobes in the lateral leaflets. These leaves are more deeply lobedthan similar unveralization leaves. E. Average number of higher orderpetiolules within medial or lateral leaflets in wild type unvernalized, AqFIEsilenced unvernalized, wild type vernalized, and AqFIE silenced leaveswith standard deviations. Both unvernalized and vernalized AqFIE silencedlateral leaflets had on average more higher order petiolules than the wildtype. Unvernalized AqFIE silenced lateral leaflets also had a slightly higheraverage number of higher order petiolules compared to wild type, butvernalized AqFIE silenced leaves had a slightly lower number of petiolulesper medial leaflet. When quantified, this increase is significant (*) atp < 0.05 for unvernalized lateral leaflets but not significant for the otherstages/leaflet types.3

Additional file 3: The PRC2 regulates carotenoid production in A.coerulea petals. A. High magnification views of epidermal cells in A. xcoerulea petal limbs. From left to right: Anthocyanin of untreated petallimb (anthocyanin is deposited in the vacuole, resulting in a very evendistribution of color), almost complete lack of color in AqANS-silencedpetal limb, and punctate pattern of carotenoid deposition in plastids ofAqEMF2-silenced petal limb. B. Expression of several A. x coeruleahomologs of genes important in carotenoid production (CRTISO and PSY)and degradation (CCD4 and NCED3) in AqANS-silenced control petals(C1-C4) and AqFIE (F1-F6) and AqEMF2 (E1 and E2) treated petals. Petalswith strong yellow pigment are highlighted in dark yellow (F1, F5, andE1) and petals with pale yellow pigment are highlighted in light yellow(F2-F4). The expression of these genes is not consistently affected in theAqFIE and AqEMF2 silenced petal samples. It is possible that other genesin the carotenoid pathway are being misexpressed. Scale bars: 10 μm.

Additional file 4: Additional candidate gene expression in PRC2VIGS-treated leaves. A. Expression of AqCLF and AqSWN in AqFIE- andAqEMF2-treated leaves. Although AqEMF2 appears to be down-regulatedin some AqFIE-silenced leaves, the expression of AqCLF and AqSWN inthese leaves is not affected. B. Expression of AqAG1, AqFL1, AqAP3-1,AqAP3-2, and AqAP3-3 in pooled AqANS silenced control leaves (C) and

AqFIE (F) and AqEMF2 (E) silenced leaves. AqAP3-1 AqAP3-2 and AqAP3-3 ismoderately up-regulated in both AqFIE and AqEMF2 silenced tissue whileAgFL1 expression is unaffected.

AbbreviationsAG: AGAMOUS; ANS: Anthocyanin synthase; AP1: APETALA1; AP3: APETALA3;Aq: Aquilegia; BP: Brevipedicellus; bp: Base pair(s); CCD: CAROTENOID CLEAVAGEDIOXYGENASE; cDNA: DNA complementary to RNA; CLF: CURLY LEAF;cm: Centimeter; CRTISO: CAROTENOID ISOMERASE; DNA: Deoxyribonucleicacid; DNase: Deoxyribonuclease; E(z): Enhancer of zeste; EMF2: EMBRYONICFLOWER 2; ESC: Extra sex combs; Eudicots: Eudicotyledonous;FIE: FERTILIZATION INDEPENDENT ENDOSPERM; FIS: FERTILIZATION INDEPENDENTSEED; FLC: FLOWERING LOCUS c; FL1: FRUITFUL-like 1; H3K27: Histone H3 Lysine27; HOX: Homeobox; IPP2: Isopentyl pyrophosphate:dimethylallylpyrophosphate isomerase; KN: KNOTTED; KNAT2: KNOTTED-like from Arabidopsisthaliana 2; KNOX: knotted1 homeobox gene; MADS: MCM1, agamous,deficiens, SRF; MEA: Medea; MKN: Moss knotted1-like; MSI1: Multi copysuppressor of IRA 1; n: Chromosome number; NCED: 9-CIS-epoxycarotenoiddioxygenase; oligo: Oligodeoxyribonucleotide; PcG: Polycomb group;Pp: Physcomitrella patens; PCR: Polymerase chain reaction; PRC1: Polycombrepressive complex1; PRC2: Polycomb repressive complex2; PSY: PhytoeneSynthase; qRT-PCR: Quantitative real time polymerase chain reaction;REF6: Relative of early flowering 6; RNA: Ribonucleic acid; RT-PCR: Reversetranscriptase PCR; SEP: SEPALLATA; STM: SHOOTMERISTEMLESS; Su(z)12: Suppressor of zeste12; SWN: SWINGER; TRV: Tobacco rattle virus;VIGS: Virus-induced gene silencing.

Competing interestsThe authors declare that they have no competing interests.

Authors’ contributionsEJG helped to conceive of the study, carried out the experiments, anddrafted the manuscript. EMK helped to conceive of the study, supervised theexperiments, and helped draft the manuscript. Both authors read andapproved the final manuscript.

AcknowledgementsThe authors would like to thank members of the Kramer lab and 2anonymous reviewers for critical comments on the manuscript.

Author details1Department of Organismic and Evolutionary Biology, Harvard University, 16Divinity Ave., Cambridge, MA 02138, USA. 2Department of Molecular andCellular Biology, Harvard University, 16 Divinity Ave., Cambridge, MA 02138,USA.

Received: 12 June 2013 Accepted: 6 November 2013Published: 20 November 2013

References1. Holliday R: Epigenetics: an overview. Dev Gen 1994, 15(6):453–457.2. Russo VEA, Martienssen R, Riggsm AD: Epigenetic Mechanisms of Gene

Regulation. Woodbury, NY, USA: Cold Spring Harbor Laboratory Press; 1996.3. Feil R: Epigenetics, an emerging discipline with broad implications. C R

Biol 2008, 331(11):837–843.4. Lewis EB: A gene complex controlling segmentation in Drosophila.

Nature 1978, 276:565–570.5. Hennig L, Derkacheva M: Diversity of Polycomb group complexes in

plants: same rules, different players? Trends Genet 2009, 25(9):414–423.6. Sawarkar R, Paro R: Interpretation of developmental signaling at

chromatin: The polycomb perspective. Dev Cell 2010, 19(5):651–661.7. Kohler C, Hennig L: Regulation of cell identity by plant Polycomb

and trithorax group proteins. Curr Opin Genet Dev 2010,20(5):541–547.

8. Schubert D, Primavesi L, Bishopp A, Roberts G, Doonan J, Jenuwein T,Goodrich J: Silencing by plant Polycomb-group genes requires dispersedtrimethylation of histone H3 at lysine 27. Embo Journal 2006,25(19):4638–4649.

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 11 of 13http://www.biomedcentral.com/1471-2229/13/185

9. Nekrasov M, Wild B, Muller J: Nucleosome binding and histonemethyltransferase activity of Drosophila PRC2. EMBO Rep 2005,6(4):348–353.

10. Pien S, Grossniklaus U: Polycomb group and trithorax group proteins inArabidopsis. Biochimica Et Biophysica Acta-Gene Structure and Expression2007, 1769(5–6):375–382.

11. Spillane C, Schmid KJ, Laoueille-Duprat S, Pien S, Escobar-Restrepo JM,Baroux C, Gagliardini V, Page DR, Wolfe KH, Grossniklaus U: Positivedarwinian selection at the imprinted MEDEA locus in plants.Nature 2007, 448(7151):349–U348.

12. Luo M, Platten D, Chaudhury A, Peacock WJ, Dennis ES: Expression,imprinting, and evolution of rice homologs of the Polycomb groupgenes. Mol Plant 2009, 2(4):711–723.

13. Whitcomb SJ, Basu A, Allis CD, Bernstein E: Polycomb Group proteins: anevolutionary perspective. Trends Genet 2007, 23:494–502.

14. Kohler C, Hennig L, Spillane C, Pien S, Gruissem W, Grossniklaus U: ThePolycomb-group protein MEDEA regulates seed development bycontrolling expression of the MADS-box gene PHERES1.Genes Dev 2003, 17(12):1540–1553.

15. Yoshida N, Yanai Y, Chen L, Kato Y, Hiratsuka J, Miwa T, Sung ZR, Takahashi S:EMBRYONIC FLOWER2, a novel polycomb group protein homolog,mediates shoot development and flowering in Arabidopsis.Plant Cell 2001, 13(11):2471–2481.

16. Gendall AR, Levy YY, Wilson A, Dean C: The VERNALIZATION2 genemediates the epigenetic regulation of vernalization in Arabidopsis.Cell 2001, 107:525–535.

17. Goodrich J, Puangsomlee P, Martin M, Long D, Meyerowitz EM, Coupland G:A polycomb-group gene regulates homeotic gene expression inArabidopsis. Nature 1997, 386(6620):44–51.

18. Oliver SN, Finnegan EJ, Dennis ES, Peacock WJ, Trevaskis B:Vernalization-induced flowering in cereals is associated with changes inhistone methylation at the VERNALIZATION1 gene. Proc Natl Acad Sci USA2009, 106(20):8386–8391.

19. Rodrigues J, Luo M, Berger F, Koltunow A: Polycomb group gene functionin sexual and asexual seed development in angiosperms.Sex Plant Reprod 2010, 23(2):123–133.

20. Okano Y, Aono N, Hiwatashi Y, Murata T, Nishiyama T, Ishikawa T, Kubo M,Hasebe M: A Polycomb Repressive Complex 2 gene regulates apogamyand gives evolutionary insights into early land plant evolution.Proc Natl Acad Sci USA 2009, 106(38):16321–16326.

21. Mosquna A, Katz A, Decker EL, Rensing SA, Reski R, Ohad N: Regulationof stem cell maintenance by the Polycomb protein FIE has beenconserved during land plant evolution. Development 2009,136(14):2433–2444.

22. Bowman J, Smyth DR, Meyerowitz E: The ABC model of flowerdevelopment: then and now. Development 2012, 139(22):4095–4098.

23. Bowman JL, Drews GN, Meyerowitz EM: Expression of the Arabidopsisfloral homeotic gene agamous is restricted to specific cell types late inflower development. Plant Cell 1991, 3:749–758.

24. Bowman JL, Smyth DR, Meyerowitz EM: Genes directing flowerdevelopment in Arabidopsis. Plant Cell 1989, 1:37–52.

25. Foronda D, de Navas LF, Garaulet DL, Sanchez-Herrero E: Function andspecificity of Hox genes. Int J Dev Biol 2009, 53(8–10):1409–1419.

26. Lopez-Vernaza M, Yang SX, Muller R, Thorpe F, de Leau E, Goodrich J:SEPALLATA3, FT and FLC genes as targets of the Polycomb group geneCURLY LEAF. PLoS ONE 2012, 7(2):e30715.

27. Wagner D: Chromatin regulation of plant development. Curr Opin PlantBiol 2003, 6(1):20–28.

28. Bharathan G, Goliber TE, Moore C, Kessler S, Pham T, Sinha NR: Homologiesin leaf form inferred from KNOXI gene expression during development.Science 2002, 296(5574):1858–1860.

29. Katz A, Oliva M, Mosquna A, Hakim O, Ohad N: FIE and CURLY LEAFpolycomb proteins interact in the regulation of homeobox geneexpression during sporophyte development. Plant Journal 2004,37(5):707–719.

30. Singer SD, Ashton NW: Revelation of ancestral roles of KNOX genes by afunctional analysis of Physcomitrella homologues. Plant Cell Reports 2007,26:2039–2054.

31. Maizel A, Busch MA, Tanahashi T, Perkovic J, Kato M, Hasebe M, Weigel D:The floral regulator LEAFY evolves by substitutions in the DNA bindingdomain. Science 2005, 308(5719):260–263.

32. Kim M, McCormick S, Timmermans M, Sinha N: The expression domain ofPHANTASTICA determines leaflet placement in compound leaves.Nature 2003, 424:438–443.

33. Hodges SA, Kramer EM: Columbines. Curr Biol 2007, 17(23):R992–R994.34. Kramer EM, Hodges SA: Aquilegia as a model system for the evolution

and ecology of petals. Philos Trans R Soc Lond B Biol Sci 2010, 365:477–490.35. Kramer EM: Aquilegia: A new model for plant development, ecology, and

evolution. Ann Rev Plant Biol 2009, 60:261–277.36. Gould B, Kramer EM: Virus-induced gene silencing as a tool for functional

analyses in the emerging model plant Aquilegia (columbine,Ranunculaceae). Plant Methods 2007, 3:6.

37. Sharma B, Kramer EM: Sub- and neofunctionalization of APETALA3paralogs have contributed to the evolution of novel floral organ identityin Aquilegia (columbine, Ranunculaceae). New Phytol 2013, 197:949–957.

38. Sharma B, Guo C, Kong H, Kramer EM: Petal-specific subfunctionalizationof an APETALA3 paralog in the Ranunculales and its implications forpetal evolution. New Phytol 2011, 190:870–883.

39. Kramer EM, Holappa L, Gould B, Jaramillo MA, Setnikov D, Santiago P:Elaboration of B gene function to include the identity of novel floralorgans in the lower eudicot Aquilegia (Ranunculaceae). Plant Cell 2007,19(3):750–766.

40. Pabón-Mora N, Sharma B, Holappa L, Kramer EM, Litt A: The AquilegiaFRUITFULL-like genes play key roles in leaf morphogenesis andinflorescence development. Plant J 2013, 74(2):197–212.

41. Gleason E, Kramer EM: Characterization of Aquilegia Polycomb RepressiveComplex 2 homologs reveals absence of imprinting. Gene 2012.

42. Ballerini ES, Kramer EM: The control of flowering time in the lowereudicot Aquilegia formosa. EvoDevo 2011, 2:4.

43. Vishnevetsky M, Ovadis M, Vainstein A: Carotenoid sequestration in plants: therole of carotenoid-associated proteins. Trends Plant Sci 1999, 4(6):232–235.

44. Ono E, Hatayama M, Isono Y, Sato T, Watanabe R, Yonekura-Sakakibara K,Fukuchi-Mizutani M, Tanaka Y, Kusumi T, Nishino T, et al: Localization of aflavonoid biosynthetic polyphenol oxidase in vacuoles.Plant J 2006, 45(2):133–143.

45. Cazzonelli CI: Carotenoids in nature: insights from plants and beyond.Funct Plant Biol 2011, 3(1):833–847.

46. Cazzonelli CI, Roberts AC, Carmody ME, Pogson BJ: Transcriptional controlof SET domain group 8 and carotenoid isomerase during Arabidopsisdevelopment. Mol Plant 2010, 3(1):174–191.

47. Ding Y, Avramova Z, Fromm M: The Arabidopsis trithorax-like factor ATX1functions in dehydration stress responses via ABA-dependent and ABA-independent pathways. Plant J 2011, 66(5):735–744.

48. Tucker SC, Hodges SA: Floral ontogeny of Aquilegia, Semiaquilegia, andIsopyrum (Ranunculaceae). Int J Plant Sci 2005, 166(4):557–574.

49. Honma T, Goto K: Complexes of MADS-box proteins are sufficient toconvert leaves into floral organs. Nature 2001, 409(6819):525–529.

50. Smaczniak C, Immink RGH, Angenent GC, Kaufmann K: Developmental andevolutionary diversity of plant MADS-domain factors: insights fromrecent studies. Development 2012, 139(17):3081–3098.

51. Immink RGH, Kaufmann K, Angenent GC: The ‘ABC’ of MADS domainprotein behaviour and interactions. Semin Cell Dev Biol 2010, 21(1):87–93.

52. Lu FL, Cui X, Zhang SB, Jenuwein T, Cao XF: Arabidopsis REF6 is a histoneH3 lysine 27 demethylase. Nat Genet 2011, 43(7):715–719.

53. Deyholos MK, Sieburth LE: Separable whorl-specific expression andnegative regulation by enhancer elements within the AGAMOUS secondintron. Plant Cell 2000, 12(10):1799–1810.

54. Hong RL, Hamaguchi L, Busch M, Weigel D: Regulatory elements of thefloral homeotic gene AGAMOUS identified by phylogenetic footprintingand shadowing. Plant Cell 2003, 15:1296–1309.

55. Hareven D, Gutfinger T, Parnis A, Eshed Y, Lifschitz E: The making of acompound leaf: genetic manipulation of leaf architecture in tomato.Cell 1996, 84(5):735–744.

56. Hay A, Barkoulas M, Tsiantis M: ASYMMETRIC LEAVES1 and auxin activitiesconverge to repress BREVIPEDICELLUS expression and promote leafdevelopment in Arabidopsis. Development 2006, 133(20):3955–3961.

57. Guo M, Thomas J, Collins G, Timmermans MCP: Direct repression of KNOXloci by the ASYMMETRIC LEAVES1 complex of Arabidopsis. Plant Cell2008, 20(1):48–58.

58. Phelps-Durr TL, Thomas J, Vahab P, Timmermans MCP: Maize roughsheath2 and its Arabidopsis orthologue ASYMMETRIC LEAVES1 interactwith HIRA, a predicted histone chaperone, to maintain knox gene

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 12 of 13http://www.biomedcentral.com/1471-2229/13/185

silencing and determinacy during organogenesis. Plant Cell 2005,17(11):2886–2898.

59. Lodha M, Marco CF, Timmermans M: The ASYMMETRIC LEAVES complexmaintains repression of KNOX homeobox genes via direct recruitment ofPolycob-repressive complex2. Genes Dev 2013, 27:596–601.

60. Lu S, Li L: Carotenoid metabolism: biosynthesis, regulation and beyond.J Integr Plant Biol 2008, 50(7):778–785.

doi:10.1186/1471-2229-13-185Cite this article as: Gleason and Kramer: Conserved roles for PolycombRepressive Complex 2 in the regulation of lateral organ development inAquilegia x coerulea ‘Origami’. BMC Plant Biology 2013 13:185.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit

Gleason and Kramer BMC Plant Biology 2013, 13:185 Page 13 of 13http://www.biomedcentral.com/1471-2229/13/185